Releases: modxcms/modAI
Releases · modxcms/modAI
0.12.0-pl
Changelog
All notable changes to this project will be documented in this file.
0.12.0-pl - 2025-05-14
🚀 Features
- Default global.text.model system setting to gpt-4o-mini (6b54015)
- Add basic tools support (5f22a3f)
- Add base model to configure agents and store conversations (#15) (2718b51)
- Add usage data to the service parsers (18fa71b)
- Separate attachments and contexts (400d6f2)
- Add tool support via DB (8d49728)
- Agent prompt & agent selector (6ca5545)
- Add support for context providers (8a83bbf)
- Add support to override model in an agent (e8a9cf2)
- Add CMP to manage context providers (bc3487b)
- Add CMP to manage tools (f582428)
- Use yes/no renderer on boolean columns (1478da8)
- Add CMP to manage agents (ddb248d)
- Ui to select agent (0bf2149)
- Add basic modx tools (30b26b1)
- Add markdown parser for assistant messages (f6d7966)
- Allow html in makrdown content (021584f)
- Persist selected agent (bbbb50b)
- Add search adn better triggers to agent/tools/cps combos (4fa8011)
- Add advanced config to the agent (2902da4)
- Add support for openrouter.ai (08c2fd9)
- Add model info to the response msg (6280e64)
- Add ACLs for client & backend (489ee77)
- Add permissions check to tools (74d2337)
- Add user groups permissions to agents (4916dd0)
- Improve scrolling for new messages (766b794)
- Add global modAI button with IndexedDB as persist layer (b1c015f)
- Use bot icon for the global button (03b8365)
- Add an option to index elements into vector db (a2d5cbb)
- Make sure user has file_create permission and save policy on source when downloading an image (1e71d3c)
- Allow to override path in download image endpoint (f29a942)
- Add indexing script (33c7dad)
- Hide agent select button if no agents are available (d1d6300)
- Update openai API to use developer instead of system, max_completion_tokens instead of max_tokens and add an ability to disable temperature (8fe8444)
- Renderer for selection context (8f28184)
- Add tooltip to the agent selector (7c9ebcb)
- Add title to the global button (0263cd5)
- Accept json encoded list of provider/tool classes (5c3631b)
- Rename tool's description to prompt and add internal description to tools & context providers (97590df)
- Allow to customize tool's prompt (3bcb3d2)
- Add a seed script (8365ef6)
- Add description to the prompt & description fields (f28031d)
- Improve chat scrolling to show max 3 lines of last user message (f2582db)
- Add EditChunk and EditTemplate tools (6e4c49b)
- Adjust image endpoint to support the new gpt-image-1 model (3761cc1)
- Make the modal resizable (b7dca8c)
- Add support for multipart/form-data endpoints like editing an image from openai (e5804df)
- Add maximize/minimize button to the header (f73c1f5)
- Store modal's position to the local storage (9f4546d)
- Define default font (7323daf)
🐛 Bug Fixes
- Remove modx dependency from composer.json (77c06bf)
- Add default usage stats for streaming (b5fd29a)
- Reenable pinecone config (9ef6dd1)
- Create/update/delete agents (78d82eb)
- Correct typo in agent prompt description (4241ada)
- Dispatch input event when setting value to the textarea (6830a6a)
- Dealy shadow's onload by 1ms to fix occasion no-scroll on first open (88c1eef)
- Prevent error saving duplicate templates and chunks (2801e3d)
- Correct prompt description to refer to editing a chunk instead of creating one (f0efe4b)
- Enforce strict array typing for runTool method arguments (4958e1a)
- Add missing lexicons (c17204d)
🚜 Refactor
- Refactor service parsers and streaming handlers (4c13dc3)
- Improve descriptions and behavior for consistency (8e22bf3)
- Update parameter 'parameters' to 'arguments' for consistency with ToolInterface (09a8734)
📚 Documentation
🎨 Styling
- New spark button (8026e60)
⚙️ Miscellaneous Tasks
- Fix commit links in the changelog (1a58beb)
- Apply phpcs fixes ([112d5bc](https://github.com/modxcm...
Contributors
types
Assets 3
0.11.0-pl
Changelog
All notable changes to this project will be documented in this file.
0.11.0-pl - 2025-03-20
🚀 Features
- Convert from processors to custom API and enable streaming when using server execution (e55a8b4)
- Abort stream on error (def1057)
- Add streaming parser for claude (960a20e)
- Add a support for local chat for free text prompts (62307df)
- Replace ext.js free text prompt with a custom ui (9b3dfe7)
- Create loading overlay on inputs while generating forced prompt (0584fcb)
- Convert generate image modal to the new UI (a712a28)
- Add an ability to switch local chat mode between text and image (f91dd91)
- Support vision from free text prompts (b0c5651)
- Add clear chat button (f27ac52)
- Add generateButton to the UI api (11f5426)
- Add getter methods for css & js files (96c57fe)
- Block closing local chat modal while generating response (41bcafd)
- Render modAI elements in shadow dom (79a9fdb)
- Streamline initing of modAI and it's security (0b5341b)
- Accept name / id as mediaSource prop for Download/Image API endpoint (1d5ac94)
- Add support for custom translation functions (fdbb562)
🐛 Bug Fixes
- Correctly propagate errors from the ai proxy (a6f5941)
📚 Documentation
🎨 Styling
⚙️ Miscellaneous Tasks
- Migrate to TypeScript (a4fd9bc)
- Migrate the resource buttons to TypeScript (31874b9)
- Update context prompt (613349b)
- Add types to localChat styles (fe4ab40)
- Add eslint and separate localChat to several files (1408c26)
- Move from webpack to esbuild (58e7125)
- Improve keyboard navigation (0c22e73)
- Add lexicons for global.image.style setting (43a60da)
- Create FUNDING.yml (836e024)
- Separate RequiredSettingException to own file (96cb930)
- Use lexicons (6b08176)
0.10.0-beta
Changelog
All notable changes to this project will be documented in this file.
0.10.0-beta - 2025-02-28
🚀 Features
- Enable calling AI services on serverside, instead of from the client (64edf9f)
- Refactor multiple cache handlers to a single history handler (a150940)
- Refactor system settings, add namespace support, add context support to the free text prompt (e6c3039)
- Add fullUrl to the return array in the image download processor (5869b40)
- Create JS API for modAI processors (b2b4b78)
- Simplify the JS API (dee659e)
- Add style option for image models (61371c3), resolves #11
- Add support for passing custom options to each model (ff25a6f)
- Merge all system prompts to a single message (314c9ca)
- Add support for client side streaming (chatgpt & gemini) (0dc2c68)
- Adjust default vision prompt (afb4cbd)
- Add cache buster for JS files (a7d4949)
🐛 Bug Fixes
- Fix return types from Settings helper class (825c33e)
- Load lexicons from processors (6587afd)
- Fix checking for empty value when getting system settings (ac4dc2b)
🚜 Refactor
- Unify fieldName as field property for processors (f90eaa7)
📚 Documentation
- Add markup and references to docs (2420b20)
- Note about how/when AI requests execute (38ef3b3)
- Describe new settings structure and streaming (b0767a0)
⚙️ Miscellaneous Tasks
0.9.0-beta
Changelog
All notable changes to this project will be documented in this file.
0.9.0-beta - 2025-02-20
🚀 Features
- Init the project (23cc847)
- Hook up to pagetitle, longtitle, description and introtext (89483e8)
- Store current field's value to the history (2f7add3)
- Generate images (dbcdd3a)
- Add quality to image generation and update baseline prompts (9716dc2)
- Hide wait msg on failure and show an error message instead (0421424)
- Adjust UI & add FreeText prompt (4ed2c92)
- Configurable tvs & resource fields (8e8ad03)
- Add setting for base output (3d88b08)
- Updating area name and consistent cross-browser modAI button styling (711113d)
- Add global.base.output (77e9379)
- Allow override base.output (093e371)
- Add support for gemini (dbc3bd0)
- Attach modAI on textareas and rte TVs (259dd34)
- Add support for anthropic models (claude) (3bea523)
- Consolidate api services (5150c47)
- Add support for custom api, compatible with openai (fb04ee6)
- Add system setting to configure download path for generated images (83c14b6)
- Add support for image generation in gemini models (83492c9)
- Add vision support for gemini (2747fa7)
- Allow model/prompt overriding for vision per field (94a2414)
- Move calling AI service to the client side (7ef59bd)
🐛 Bug Fixes
- Add missing import (90a2f66)
- Update assetsUrl path (68780fa)
- Fix saving altTag (9e62f83), Resolves #1
- Fix name of global base prompt (a68bf91)
- Fix global base prompt in free text (faaf61d)
- Fix grabbing chatgpt key (a173a55)
- Disable timeout when calling prompt processors (56b33d8)
- Fix return of generateImage from custom chatgpt integration (8539f35)
- Use getImageFieldSetting when generating an image (95f5d90)
📚 Documentation
- More descriptions and initial documentation (3cefccf)
- Document base prompt (e100400)
- Update README.md (c2e4e17)
- Fixing readme typos and clarifications (758349f)
- Streamline README (deaee2f)
- Clarify the README.md even more, because I can't stop (62ae55d)
- Update instructions for TV handling + default image to wide (634e8e0)
- Create LICENSE (5f24386)
- Move readme (d65c986)
- Documentation updates and Settings names (808b6d4)
- Multi-model and custom model usage instructions (97aa5a0)
- Fix typos in README (a89293b)
- Adjust order and add service specific docs (0700b38)
- Add Gemini Image + Vision (709bd5d)
- Convert readme to writerside (fb240bd)
- Set docs to build while in private mode (e2ccdc7)
- Fix link to docs and add theme (a13cdc0)
🎨 Styling
- CSS for nav buttons (e116c94)
- Make buttons great again (76a1f04)
- Initial tweaks for Image+ alt button (still need 3px or so top margin) (ed89a1d)
- Fix for minor alignment on alt text icon (1f3afb6)
- Adjust AI button look (e7e4929)
- Increase height of response textarea in the free text prompt window (c7533da)
- Visually simplify button borders (1b48a84)